WO2013093393A1

WO2013093393A1 - Method and apparatus for an on-channel repeater

Info

Publication number: WO2013093393A1
Application number: PCT/GB2012/000909
Authority: WO
Inventors: Peter Moss; Adam WIEWIEORKA
Original assignee: British Broadcasting Corporation
Priority date: 2011-12-19
Filing date: 2012-12-17
Publication date: 2013-06-27
Also published as: GB201121951D0; GB2497772B; GB2497772A

Abstract

An on-channel repeater having a receiving antenna, for receiving an RF signal to produce a received signal, a plurality of transmitting antennas, each for transmitting a respective transmitted signal, and a plurality of inputs, each for receiving a respective reference signal from each respective transmitted signal is provided. The transmitted signal comprises an amplified recovered signal derived from the received signal as well as a respective potentially interfering signal that may produce interference at the frequency of the received signal. A plurality of adaptive filters each coupled to a respective reference signal input are provided and controlled by a plurality of control coefficients to produce respective cancellation signals. A combiner then combines the cancellation and received signals to produce the recovered signal. A power amplifier is arranged to receive, amplify and transmit the recovered signal on each of the transmitting antennas.

Description

METHOD AND APPARATUS FOR AN ON-CHANNEL REPEATER

BACKGROUND OF THE INVENTION

This invention relates to rebroadcast transceivers which are designed to receive RF (radio frequency) signals, amplify them, and retransmit them onward on the same frequency. Such transceivers are known in the broadcasting field as on- channel repeaters, and are sometimes termed active deflectors.

In an on-channel repeater, due to unwanted coupling or feedback between the receiving and the transmitting antenna, the repeater can also receive its own retransmitted output, thus causing instability and relaxation oscillations. Our International Patent Application W097/14642 and European Patent EP 0772310 describe a method and apparatus which has been found to be surprisingly effective in removing this feedback in systems carrying noise-like signals. In this method there is an amplification path between the input and output antennas which provides substantially linear processing and includes a delay sufficient to decorrelate the output and input. The repeater includes an amplification path providing substantially linear processing without demodulation and decoding, and a filter-estimator responsive to the signal in the amplification path for correlating the signal in the amplification path before the delay with a signal taken after the delay to produce a plurality of correlation coefficients. The filter estimator may use the least mean square method. An adaptive filter in the form of a transversal filter receives the signal in the amplification path after the delay and is controlled by the control coefficients to provide a cancellation signal, and a combiner that combines the cancellation signal with the signal in the amplification path so as to reduce the effect of the feedback. In this way, unwanted feedback from the output of the active deflector to the input is substantially eliminated. The compensation conveniently makes use of the inherent noise-like property of the signal, as described it is an OFDM signal; however a separate noise signal may be added if necessary. In our previous application EP1724946 we described a method of reducing the impact of co-channel interference originating from a neighbouring transmission on the same site deriving the reference signal for the filter estimator and the transversal filter from a point after the combining of the neighbouring transmissions and the repeater's own output.

SUMMARY OF THE INVENTION

We have appreciated the need to improve on-channel repeaters in circumstances where the on-channel repeater is to be deployed co-sited with multiple, distinct, potentially interfering transmissions. In particular, we have appreciated the need for improvements in situations where the interferer is intermittent, pulsed or muted for extended periods. The invention is defined in the claims to which reference is now directed.

An embodiment of the invention derives separate cancellation signals for each respective potentially interfering transmission and combines these cancellation signals with a received signal to produce a recovered signal. The cancellation signals are derived using a filter estimator and an adaptive filter using correlation, thereby, extracting a linearly dependent part of the signal. In this way, such unwanted signals that are linearly dependent on respective reference signals may be removed. By subtracting such cancellation signals prior to a single power amplification stage, the arrangement only requires an adaptive filter for each separate potentially interfering signal.

A preferred embodiment includes a management unit for controlling the coefficients of the filters on detecting the presence or absence of one of more of the potentially interfering signals. BRIEF DESCRIPTION OF THE DRAWINGS

The invention will now be described in more detail by way of example with reference to the drawings, in which: is a block diagram of first known on-channel repeater; is a block diagram of a second known on-channel repeater;

is a block diagram of a finite impulse response filter; shows one tap of a multitap architecture; and

is a block diagram of an on channel repeater embodying the invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

The invention may be embodied in a method for processing signals, a filter arrangement and also to an on-channel repeater, echo, noise or interference canceller or control system. To ease understanding, existing systems will first be described to explain the signal nomenclature with reference to the systems already mentioned and as shown in Figures 1 to 4.

An example of a transceiver of the type described in our earlier applications is shown in Figure 1. The transceiver 10 has a receiving antenna 12 which is coupled to an amplification path 14 which includes an adder 16, a decorrelating delay 18, and an amplifier 20. The output of the amplifier is applied to a transmitting antenna 22. The output of the delay 18 is also applied as a reference signal to a filter estimator 24, which also receives the output of adder 16, and an adaptive filter 26, which applies an output to the subtractive or inverting input of adder 16. The construction and operation of the whole of this corrector circuitry 10 is described in detail in our earlier applications. Part of the output of the transmitter antenna will be picked up by the receiving antenna as indicated by the dashed line 30, as unwanted feedback. The corrector 10 removes the effect of this feedback. An on-channel repeater can be used for two purposes, namely coverage extension and hole filling. Coverage extension is used if the received signal strength is insufficient in a particular area, and the repeater is then used as an additional transmitter, often as a part of a Single Frequency Network (SFN).

In hole-filling, the received signal strength is sufficient, but signals on adjacent channels to the desired signal are so strong that domestic receivers have insufficient dynamic range to demodulate the wanted channel successfully. In extreme cases, in areas very close to adjacent channel transmitters, the intermodulation products from the adjacent channels interfere directly with the wanted signal, even though spectral mask requirements have been fulfilled. The repeater must remove the intermodulation products of co-sited adjacent channels that arrive at its input, as well as removing the unwanted feedback. Our earlier European Patent EP 1724946 describes an improved on channel repeater arrangement as shown in Figure 2. In this arrangement, the on channel repeater has a feedback signal that enables cancellation of any uncorrelated parasitic signal arriving at the receiving antenna which was present in the transmit antenna feed, not just the recovered version of the wanted signal. In order to avoid instability caused by parasitic coupling between the receiving and transmitting antennas, stray transmitted signals are cancelled by an adaptive filter. Whereas previously this filter has processed the retransmitted signal obtained internally in the repeater and tried to model the external parasitic feedback, this arrangement uses external feedback from the transmitting antenna feed, that includes any distortion, intermodulation products and interference from adjacent channels.

The arrangement of Figure 2 has similar features to that of Figure 1. The transceiver 10 has a receiving antenna 12 which is coupled to an amplification path 14 which includes an adder 16, a decorrelating delay 18, and an amplifier 20. The output of the amplifier is applied to a transmitting antenna 22. A reference signal is applied to a filter estimator 24, which also receives the output of adder 16, and an adaptive filter 26, which applies an output to the subtractive or inverting input of adder 16. The filter estimator 24 generates filter coefficients which are applied to the adaptive filter 26.

The differences from Figure 1 are first that in this case the reference signal, that is the input to the adaptive filter 26 and the reference input to the filter estimator 24, are taken not from within the transceiver but rather from the transmitter antenna 22, as indicated at 40. That is to say, the reference signal is taken after the transmitter power amplifier. Secondly, the potentially interfering signal on the adjacent channel, indicated at 42, is arranged to be transmitted on the same transmitting antenna as the transmitting antenna 22 used by the repeater, by virtue of the combiner 44. The combiner may in fact be a coupler, or may be absent altogether if no adjacent channel transmitter is present. The system nevertheless still mitigates the effects of power amplifier intermodulation on the feedback cancellation process. If the repeater is co-sited with one or more adjacent channel transmitters, by transmitting from the same antenna as the adjacent channel or channels, the reference input of the repeater can then readily be fed with a combined signal transmitted through the antenna that contains not only the wanted but also the adjacent channel. In this way not only can the principal linear feedback of the repeated channel be removed, but also the intermodulation products which are present in the interfering signal can be cancelled as well as those caused by the retransmission of the wanted signal, avoiding undesirable re-radiation.

To summarise, and to explain the nomenclature of the signals in Figure 2, an on channel repeater uses an adaptive finite impulse response (FIR) filter h 26 to extract a signal ε(ί) referred to as a recovered signal, that is as similar as possible to an original wanted signal s(t), from a received corrupted input signal x(t). The objective of the repeater 10 is to make the recovered signal ε(ι) as similar as possible to the original wanted signal s(t). Then the recovered signal ε(ϊ) is amplified for retransmission. In order to derive the recovered signal e(t) a cancellation signal z(t) is derived using an adaptive filter 26 in accordance with coefficients provided by a filter estimator 24. The cancellation signal is subtracted from the input signal x(t) using a combiner 16 to produce the recovered signal e(t). The filter estimator is provided with the recovered signal e(t) and also a reference signal y(f), which is itself derived from ε(() in a linear manner and also contains any in band interference from an adjacent channel 42 mixed in a combiner 44. By using a single antenna arrangement for transmitting the amplified recovered signal t(t) and a potentially interfering signal, the reference signal y(t) accurately represents the potential interference from the retransmitted signal and the potentially interfering signal as previously described. The operation of the adaptive filter and filter estimator may be in accordance with our earlier patents such as EP 0772310.

For completeness, and to aid understanding of the equations discussed in relation to an embodiment below, an example finite impulse response (FIR) filter and the manner in which coefficients for that filter may be derived will now be described with reference to Figures 3 and 4. As is known to the skilled person, the output of a FIR is a weighted sum of the current and a finite number of previous values of the input.

The operation is described by the following equation, which defines the output sequence in terms of its input sequence:

2(t) = o y(t) + h, y(t-1) + h₂ y(t-2) + ....+ h_Ny(t-N) which may be re-written as

Where: y(t) is the input signal at time sequence t

z(t) is the cancellation signal at time sequence t

hj are the filter coefficients (also known as tap weights)

N is the filter order (also referred to as the number of taps) The properties if such a filter are defined by the choice of coefficients h, which define how samples at a give point in time are summed. Such an example is shown in Figure 3. The filter estimator 24 produces coefficients for the FIR filter by performing correlations between the reference signal y(t) and the recovered signal ε (t). One element of the correlator bank used for determining coefficients is shown in Figure 4. As is shown, the tap estimator looks for the presence of the output reference signal y(t) in the recovered signal s(t) . The reference signal y(t) is decorrelated from ε(ί) by the decorrelating delay, though, and so this signal is not found. Instead, the interfering signal is found. The recovered signal ε(£) is sometimes referred to in literature as the "error" signal as such filters are used to minimise this signal. In the present arrangement, though, ε (t) is the signal that we wish to amplify and so is referred to as the recovered signal. The output of each correlator h is the corresponding coefficient used for a corresponding tap of the FIR filter shown in Figure 3. A full explanation of this may be found in our earlier patent EP 0772310.

We have appreciated the need in particular to reduce interference between signals transmitted from cellular base station transmitters (referred to an electronic communications network or ECN) and digital terrestrial television (DTT) signals. Under current proposals, the radio spectrum is such that DTT signals may suffer interference from ECN base stations such that a TV receiver may fail to operate. This scenario is referred to as "hole punching". Such issues are discussed in a paper entitled "The Feasibility of DVB-T on channel repeaters for coverage repair on Channel 60", Ofcom 2109/OCR/R/3.0, dated 5 August 2009. In that paper, it is noted that the use of on channel repeaters using the same antennas for both a mobile telephone ECN and digital television DTT service has a number of benefits. Use of the same antenna ensures that the difference in field strength between the two services will remain as constant as possible. In addition, using the same antenna allows the possibility to cancel intermodulation products from the ECN signal that fall in the DTT channel as already described in relation to Figure 2. We have appreciated, however, a problem with such an arrangement for an on channel repeater located at the same location as a mobile telephone base station antenna. Typically, a mobile network comprises cells with each cell being divided into three sectors with a separate antenna directed to each sector which in the proposed mode of operation requires that each sector re-radiates the on-channel repeater's output signal. Accordingly, it is necessary to cancel the unwanted feedback resulting from these three separate signals. The slight difference in antenna location renders the cancellation of any unwanted interfering signals non trivial.

In the light of the potential need to provide 'network repair' functionality in the context of LTE (Long Term Evolution) 4G wireless roll-out, where base-stations may have three or more sectors with independent transmissions, we have appreciated that improvement is needed.

In general we have appreciated the need for an on-channel repeater to provide amplification of a received signal whilst avoiding amplifying unwanted feedback and also reducing the impact of interfering signals. The example of three sectors as will be used in the context of LTE is one example. This may be extended further in having multiple antennas serving each of three sectors, such as four antennas serving each of three sectors giving twelve potential interfering signals. More generally, a plurality of possible interfering signals may be present. An extension of the scheme of Figures 1 to 4 is proposed as illustrated in the embodiment of Figure 5.

Figure 5 shows how multiple outputs each comprising the wanted OFDM output together with (for example) adjacent (different) LTE signals may be configured to provide three distinct reference inputs to a revised OCR which still features a single through path and decorrelating delay but is able to cancel feedback present at the input from all N transmission antennas simultaneously. The nature of the design allows the interferers to have an intermittent or pulsed nature without degrading system performance unacceptably. In addition stable operation is maintained even if one or more adjacent channel transmitters are muted.

The system as a whole comprises an input antenna 12 arranged to receive a signal that is to be retransmitted. In the case of a terrestrial television on-channel repeater, the input antenna 12 may be a directional antenna directed towards the source of the television transmission. The on-channel repeater has a plurality of transmitting antennas 22, 122, 222 for transmitting the amplified received signal, as well as transmitting another service in this case separate LTE signals to be broadcast in different sectors.

The arrangement in Figure 5 receives a signal containing a wanted signal s(t) which is provided on line 12 (with appropriate RF mixing and analogue to digital conversion) as digital signal x(t), referred to as a received signal. A combiner 16 receives the now digitised received signal x(t) on a non inverting input and three separate signals to be removed on an inverting input. The three separate signals z1 (t), z2(t) and z3(t), referred to as cancellation signals, are provided on lines 127, 227, 327 from respective adaptive filters 126, 226, 326. The cancellation signals are summed within the combiner to give z(t) and subtracted from the received signal x(t) giving the output of the combiner as the an output signal ε(ί), referred to as the recovered signal. The recovered signal ε(ί) is provided to a decorrelating delay 18 and converted to an analogue signal prior to an input of a power amplifier 20. The output of the amplifier is therefore an amplified recovered signal and is provided to three separate combiners 144, 244, 344 each receiving a respective mobile telephone signal to be transmitted as a separate sector within a mobile telephone cell. The mobile telephone signals may each potentially interfer with the received signal at the frequency of the received signal and so may be referred to as potentially interfering signals. Typically, this will be because the mobile telephone signals are on an adjacent channel, as described above, and so may have energy at the frequency of the recived signal. The combiners provide the combined recovered signals and each separate mobile telephone signal and provide these to respective antennas for transmission as transmitted signals which, as shown in the diagram, may also provide unwanted feedback to the receiving antenna. A reference signal is also taken from the feed to each antenna, digitised and provided as digital reference signals y1 (t), y2(t) and y3(t) to the respective filter estimators 124, 224, 324 and adaptive filters 26, 226 and 326. The reference signals are from the signals to be transmitted in the sense that they are substantially the same as the signals to be transmitted, other than for any changes needed for processing such as filtering and analogue to digital conversion. The reference signals are taken from the feed to each antenna, which may be by some form of coupling, but could be taken in othe ways including directly from the antennas themselves. In addition, the recovered signal ε(ή is provided at a separate input to each respective filter estimator. In this way, each separate filter estimator is provided with a reference signal containing the corresponding potentially interfering signal. The filter estimator derives coefficients for the corresponding adaptive filter in the manner previously described.

It is noted that the arrangement of Figure 5 does not simply comprise three separate unrelated filtering channels. Instead, the cancellation signals are combined to give a total cancellation signal z(t) and combined by being subtracted from the received signal x(t) prior to a common decorrelating delay that delays the output recovered signal ε(/) prior to amplification. The decorrelating delay 18 is shown schematically as a separate component but may, of course, comprise delay inherent within the processing circuitry rather than providing an explicit separate delay component. An advantage of this approach is that it accommodates for the fact that the LTE signals are not sufficiently directional for there to be no overlap between them. By subtracting the cancellation signals prior to a single power amplification stage 20, the arrangement only requires an adaptive filter for each separate potentially interfering signal to be broadcast alongside the received signal that is to be amplified.

A problem exists in knowing how to derive appropriate coefficients in certain circumstances. In particular, the potential interferers are not constant signals but at any time a given potential interferer may be silent or, indeed, a given sector may not be used for a long period. In such circumstances, the filter estimators may incorrectly set the corresponding filters. For this purpose, a filter management unit 15 is provided and coupled to each of the three input potentially interfering signals and also to each respective filter estimator. The architecture shown in Figure 5, in particular the multiple summation of cancelling filter outputs to provide a single "error" term in the form of the recovered signal ε( and a combined cancellation signal z(t) is instrumental in providing stable operation. The filter management block 15 influences filter behaviour during periods when one or more adjacent channel transmissions are absent. It does this by applying one or more of the processes described below to the subset of outputs which have no adjacent signal. The processes may be used alone or in various combinations. The filter management unit performs the function of controlling the filter coefficients produced by the filter estimators in the event that the absence of one or more of the potentially interfering signals is detected. In order to monitor for the presence or absence of such interfering signals, each of the sources of the separate LTE signals 142, 242, 342 is coupled by a respective line 143, 243, 343 to a respective input of the filter management unit. In one approach, the signal itself may be provided to the filter management unit, which then monitors for the presence or absence of the signal. As an alternative, each of the sources may include functionality to assert an indicator on the respective output lines 143, 243, 343 indicating the presence or absence of the corresponding LTE signal. Whatever approach is chosen, the filter management unit 15 receives an indication as to whether each potential interfering signal is present.

On detecting that only one LTE signal is no longer present, the filter management unit can allow each of the filter estimators to continue without altering their behaviour. The reason for this can be seen by considering the reference signals provided to each of the filter estimators. If the output of the power amplifier is represented as C and each of the in band interfering signals as LTEi, LTE₂ and LTE₃ respectively. Y^t) = C + LTE^

Y₂(t) = C + LTE₂

Y₃(t) = C + LTE₃ If we now consider that, say, the signal LTE₃ is no longer present, it can be seen that Yi(l), Y₂(t) and Y₃(t) remain as three different signals. When each of the filter estimators operate, there is a unique solution to producing the cancellation signals z1 (t), z2(t) and z3(t). However, if two or more of the LTE signals are no longer present, it is no longer the case that the filter estimators may find a unique solution. For example, if signals LTE₂ and LTE₃ are no longer present, this means the inputs to the second and third estimators simply become the common signal C. As a result, the second and third filter estimators are both attempting to produce the same cancellation signal. As the outputs are connected in the same signal path, the result is that the filter estimators may be unstable with one drifting in a positive direction and one a negative direction, both attempting to counteract each other. On detecting that two or more of the potentially interfering signals to be transmitted on the same frequency as the input signal are no longer present, the filter management unit takes control of the respective filter estimators. For the avoidance of doubt, the filter management unit and the filter estimators may comprise separate functional units or may be combined in a single functional unit.

In the following explanation of processes e(k) is the error, that is the recovered signal, at time index k and y_L (k) is the vector of reference signal samples at the same time at the input to the L_th adaptive filter, λ and μ are constants determining convergence rate, typically within the range 0 to 0.1 .

On detecting that two or more of the signals to be transmitted on the same frequency are no longer present, as a first step, the subset of filters for which no interfering signal, such as an LTE signal, is present are determined. The subset of filters currently without LTE signal excitation is defined using vectors to define the coefficients as:

(h, h₂ h₃ h _v } (1 )

The filters may be considered as vectors as each comprises multiple coefficients and the filters having the same number of filter taps. The filters in the subset of filters having no interfering signal are therefore operating on the same signal as each other (the common signal C found above) and so these are properly represented as vectors.

As already noted, in the absence of any potentially interfering signal all y_L(k) are nominally equal within the subset of filters (as there will be no excitation by any potentially interfering LTE signal).

The filters in the subset N of filters having no LTE signal are then updated in accordance with one or more of the following processes as follows.

Process 1

Cause gradual updating of each filter towards the average value of the filters, starting at the value existing just before cessation of the adjacent channel signals, but allowing the average to move in response to the changing cancellation requirement.

Define a target average tap vector of the filters in the subset N using the values of the taps at time index k=0, defined as the moment the subset of filters without LTE excitation is established:

Update the target average to address changing feedback requirements using conventional L S algorithm:

By way of explanation of equation 3, the signal Y„ is the same for all filters in the subset, as already mentioned, and so by summing the LMS for the average filter, the average will track the feedback requirement for all filters. In addition, the process updates each filter towards the target average: h_e(k + l) = b_n (k) + (k) - h_n (kj) * e {l .Jv-} (4)

Using this process, the coefficients of the filters within the subset N are directly controlled so as to gradually depart from their separate values at the point at which the LTE signals were no longer present and converge towards an average vector. As an example, if each of the filters has just two coefficients, then the first coefficient of the filters in the subset are summed together and divided by the number of filters in the subset. The second coefficient of each filter are summed together and again divided by the number of filters in the subset. These values then give the average filter vector towards which each filter should converge, thereby explicitly avoiding the problem of tap divergence.

An advantage of this process is that the filters gradually converge and in doing so reduce the chance of glitches due to sudden changes. One possible disadvantage, though, is that when the LTE signals are present again and the filter management unit hands control back to the filter estimators, the filter estimators are starting from the average vector value and not near the positions that they should be in for effective cancelling. A potential improvement upon this approach is to use the process of Process 2 set out below.

Process 2

This process is to freeze all filter coefficients of the filters in the subset, except for the coefficients of one filter in the subset, to their last-known value, allowing the remaining one filter in the subset to track the changing cancellation requirement and to continue to update its coefficients.

Suppose only the i^ih filter is updated: h, (* + l) = h, (*) + A£(*)y; (*) (5) The remainder are frozen: h„(* + l) = h„(*) _n≠i (6) An advantage of Process 2 is that it allows just one filter in the subset to do the work of tracking changes, so that all other filters in the subset maintain the filter coefficients that were appropriate in the presence of their corresponding LTE signals. As a result, when the LTE signals resume, the filters that were frozen may already be in appropriate positions (with appropriate coefficients). This approach will improve the speed at which the filters can resume tracking and cancelling interfering signals when they resume. In scenarios where there is not much variation due to external factors such as atmospheric condition or changing geographical environments, the coefficient values as they were frozen at the point that the LTE signals were lost, may in fact be appropriate for when the LTE signals resume.

Process 3

Starting with each filter at its last-known values, allow all filters to update by a common amount to track the changing cancellation requirement.

All filters update: h_n(k + \) = „(k) + (k)y (k) n {l...N} ... . (7)

In this approach, the filters in the subset are effectively treated as a single large filter.

A variant of this option allows for different convergence factors on each filter and tap:

(^k + !) = K (*) ^ ~^k y_n ^' _q (k) n≡{\...N}q {\.. (7A) where Q is the number of coefficients of the filters. As an example of the use of this variant, the value of X_nq can be made proportional to the magnitude of h_nq at time index k=0.

This variant option allows different rates to be defined at which the filters are updated on a tap-by-tap basis.

Process 4

A combination of processes (2) and (3) whereby a subset of the subset of muted outputs is frozen, allowing the remaining filters to share the task of tracking the changing cancellation requirement. Label the frozen sub-subset (without loss of generality) h₂ h h, } j < N (8)

And the tracking (unfrozen) sub-subset

Then h_n(* + l) = h„(A;) n e {I-/} (10) and h„(* + l) = h_B (*) + -— s(*)y;(*) « e {y + l...N} .(1 1)

N - j This last process may be particularly beneficial in implementations having larger numbers of separate in band interfering signals, such as the example of three LTE signals with four separate MI O signals giving a total of twelve separate antennas and separate interfering signals. In this example, if, say, six of the LTE signals were no longer present, four of these six could be frozen and the remaining two operate together, in accordance with Process 3 to attract the changing cancellation requirement.

Yet further variations are possible, such as saving all of the filter coefficient values at the point that a corresponding LTE signal ceases and then operating one of the processes discussed above during the time that the LTE signal is not present. On the LTE signal resuming, the saved values may be retrieved and used again as initial values prior to recommencing tracking using the processes discussed herein.

An additional stability management procedure may also be included: If a gradual complementary divergence of the kth tap of a pair or n-tuple of filters is detected then this is corrected by introducing a small bias to each of the kth taps or by a 'hard reset' of tap values such as to restore the tap vectors to within acceptable limits. All such changes are carried out in such a way that the sum of the kth tap coefficients across all filters is held constant thereby ensuring uninterrupted and effective feedback cancellation performance.

Claims

1. An on-channel repeater, comprising:

a receiving antenna (12) for receiving an RF signal to produce a received signal;

a plurality of transmitting antennas (22, 122, 222) each for transmitting a respective transmitted signal comprising an amplified recovered signal derived from the received signal as well as a respective potentially interfering signal that may produce interference at the frequency of the received signal;

a plurality of inputs (40, 140, 240) each for receiving a respective reference signal from each respective transmitted signal;

means (124, 224, 324) coupled to each respective reference signal input for producing a plurality of control coefficients;

a plurality of adaptive filters (126, 226, 326) each coupled to a respective reference signal input and controlled by the control coefficients to provide a respective cancellation signal;

a combiner (16) for combining the cancellation signals with the received signal to produce the recovered signal;

a power amplifier arranged to receive the recovered signal to provide an amplified recovered signal and apply it to each of the transmitting antennas for transmission.

2. An on-channel repeater according to claim 1, further comprising a filter management unit arranged to detect the presence or absence of each of the respective potentially interfering signals for transmission and to control at least some of the coefficients.

3. An on-channel repeater according to claim 2, wherein the filter management unit is arranged to control the coefficients of those filters in a subset of filters for which a corresponding potentially interfering signal for transmission is determined to be absent.

4. An on-channel repeater according to claim 3, wherein the filter management unit is arranged to control the coefficients of those filters in the subset by updating each filter in the subset towards the average value of the filters in the subset.

5. An on-channel repeater according to claim 3, wherein the filter management unit is arranged to control the coefficients of those filters in the subset by freezing the coefficients of at least some of the filters in the subset.

6. An on-channel repeater according to claim 3, wherein the filter management unit allows the coefficients of the remaining filters in the subset to continue to be updated.

7. An on-channel repeater according to claim 6, wherein the coefficients of the remaining filters in the subset are controlled by the filter management unit such that they are updated by a common amount.

8. An on-channel repeater according to claim 2, wherein the filter management unit is arranged to control the coefficients of those filters in the subset by updating the coefficients of all filters in the subset by a common amount.

9. An on-channel repeater according to any preceding claim, in which the coefficient producing means (124, 224, 324) receives the respective reference signal as one input and the recovered signal as another.

10. An on-channel repeater according to any preceding claim, in which the on-channel repeater is operative with OFDM signals.

11. An on-channel repeater according to any preceding claim, in which the on-channel repeater is a digital television repeater.

12. An on-channel repeater according to any preceding claim, wherein the respective potentially interfering signals that may produce interference are neighbouring or adjacent channels.

13. A on channel repeater according to claim 12, wherein the respective potentially interfering signals that may produce interference are LTE signals, each respective signal being for a separate cell sector.

14. An on channel repeater according to any preceding claim wherein the received signal is a digital television signal.

15. A transmitter arrangement comprising an on-channel repeater according to any preceding claim and transmitters which transmits potentially interfering signals, the output of the on-channel repeater and outputs of the transmitters being coupled to the said transmitter antennas.

16. An method, comprising:

producing a received signal from a receiving antenna ( 2);

transmitting a respective transmitted signal comprising an amplified recovered signal derived from the received signal as well as a respective potentially interfering signal that may produce interference at the frequency of the received signal from each of a plurality of transmitting antennas (22, 122, 222); receiving a respective reference signal from each respective transmitted signal at a plurality of inputs (40, 140, 240);

producing a plurality of control coefficients using each respective reference signal input;

providing a respective cancellation signal from each of a plurality of adaptive filters (126, 226, 326) each coupled to a respective reference signal input and controlled by the control coefficients;

combining the cancellation signals with the received signal to produce the recovered signal;

amplifying the recovered signal to provide an amplified recovered signal and applying it to each of the transmitting antennas for transmission.

17. A method according to claim 16, further comprising detecting the presence or absence of each of the respective potentially interfering signals for transmission and controlling at least some of the coefficients.

18. A method according to claim 17, comprising controlling the coefficients of those filters in a subset of filters for which a corresponding potentially interfering signal for transmission is determined to be absent.

19. A method according to claim 17, comprising controlling the coefficients of those filters in the subset by updating each filter in the subset towards the average value of the filters in the subset.

20. A method according to claim 17, comprising controlling the coefficients of those filters in the subset by freezing the coefficients of at least some of the filters in the subset.

21. A method according to claim 7, comprising allowing the coefficients of the remaining filters in the subset to continue to be updated.

22. A method according to claim 21 , comprising controlling the coefficients of the remaining filters in the subset such that they are updated by a common amount.

23. A method according to claim 17, comprising controlling the coefficients of those filters in the subset by updating the coefficients of all filters in the subset by a common amount.